Randomized Kaczmarz with Averaging
Numerical Analysis
2020-02-12 v1 Numerical Analysis
Abstract
The randomized Kaczmarz (RK) method is an iterative method for approximating the least-squares solution of large linear systems of equations. The standard RK method uses sequential updates, making parallel computation difficult. Here, we study a parallel version of RK where a weighted average of independent updates is used. We analyze the convergence of RK with averaging and demonstrate its performance empirically. We show that as the number of threads increases, the rate of convergence improves and the convergence horizon for inconsistent systems decreases.
Cite
@article{arxiv.2002.04126,
title = {Randomized Kaczmarz with Averaging},
author = {Jacob D. Moorman and Thomas K. Tu and Denali Molitor and Deanna Needell},
journal= {arXiv preprint arXiv:2002.04126},
year = {2020}
}
Comments
19 pages, 9 figures